Microsoft's new AI tool can create a GIF out of any image

A new artificial intelligence model called Pix2Gif is introduced by Microsoft's research group. It can convert a still image into a gif in a couple of seconds. Pix2Gif employs an "image translation" method while utilizing the same kind of diffusion model as many AI text-to-video models, such as Runway and Pika Labs. It is a method of creating AI that combines 3D models, video, and images.

As per the researchers, the users need to guide the model about the motion through a text prompt along with the image input. With this, the tool "spatially transform" the features of the original image based on the text prompt.    

To create a two-second gif from a still image, the AI tool takes nearly a minute. However, the tool may create GIFs even quicker with a faster GPU. In order to train the model, the researchers claimed that they have created 100,000 animated gifs with appropriate captions. After that, they extracted frames from the gifs and used the captions as the text prompt.  

Microsoft may not make this new tool public as this is more of a research project and make its way into Copilot as a standalone tool, as per report.

It is available for everyone in the test environment. The users can provide an image to the tool or a text prompt and get a GIF. As per the report, Microsoft may include this tool in Designer or Paint to make it easier to animate and may even make adjustments to an image using AI.